[
https://issues.apache.org/jira/browse/BEAM-13715?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17480209#comment-17480209
]
Kenneth Knowles commented on BEAM-13715:
----------------------------------------
CC [~boyuanz] (author) [~aromanenko] (reviewer) for
https://github.com/apache/beam/pull/12572
> Kafka commit offset drop data on failure for runners that have
> non-checkpointing shuffle
> ----------------------------------------------------------------------------------------
>
> Key: BEAM-13715
> URL: https://issues.apache.org/jira/browse/BEAM-13715
> Project: Beam
> Issue Type: Bug
> Components: io-java-kafka
> Affects Versions: 2.26.0, 2.27.0, 2.28.0, 2.29.0, 2.30.0, 2.31.0, 2.32.0,
> 2.33.0, 2.34.0, 2.35.0, 2.36.0
> Reporter: Kenneth Knowles
> Priority: P1
>
> A user noticed that we commit Kafka offsets without any obvious
> checkpointing. We use a {{Reshuffle.byRandomKey()}} to cause Dataflow and the
> SparkRunner to checkpoint. But on runners with non-checkpointing shuffle,
> this risks data loss.
>
> The modern solution is to use {{{}@RequiresStableInput{}}}. This is not
> perfectly/fully implemented across many runners, so we still need the
> explicit shuffle for now.
>
> https://stackoverflow.com/questions/70785672/apache-beam-kafkaio-commit-offset-behaviour
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
